qrian

8 ms·

31.
by qrian 3y ago
nope, all my prompts lead the the same answer.
32.
by qrian 3y ago
yes, I loaded the 1.35GB .bin file.
33.
by qrian 3y ago
It only says 'affor affor affor...(repeated 100+ times)' for me. I am not sure how to debug this.
34.
by qrian 3y ago
[Docters still use pagers]( https://ferd.ca/notes/rhip-doctors-and-pagers.html ) and [put password in post-it notes on the device]( https://ferd.ca/notes/paper-you-want-my-password-or-a-dead-p... ) no
35.
by qrian 3y ago
I have heard 'bullshitting' as a term for this. Be underhandedly deceptive that is. But I do have to say a lot of humans also bullshit throught their works and get away with it, so I don't know if this problems is fixable.
36.
by qrian 3y ago
Looks like you are correct. I did not know about directed angles. Thanks.
37.
by qrian 3y ago
But then again, there's an error in the proof of IMO P3, in Fig1.f and Step 26. of the full proof in supplimentary material[1], which it states that ∠GMD = ∠GO2D, which is incorrect. It should be ∠GMD + ∠GO2D = π. I am trying to follow
38.
by qrian 3y ago
I was all ready to be skeptical for most of these kind of works its outputs are 'not like human proofs' but then I saw Evan Chen's quote that it is indeed clean human-readable proof. Evan Chen is a prominent member of Olympia
39.
by qrian 3y ago
The high school math Olympiad team is the fondest memory of mine. It was like solving puzzle games every day. But I guess if you are forced into puzzle games when you don’t like it can be a torture. I also changed my career to tech but I st
40.
by qrian 3y ago
I've seen models learn heuristics that are harmful in real performance, and I wonder how much is accuracy directly transferrable to actually good drafting. A question, when GPT-4 contradicts in explanation, how much of them were in fac
41.
by qrian 3y ago
Seems pretty cool. I've heard great things about deep brain stimulation and how it is the SOTA for many neurological disorders. It has actual experiments in humans instead of rodents. And the humans are healthy adults too! So it tick
42.
ChatGPT cut off date now April 2023(chat.openai.com)
112 points by qrian 3y ago|87 comments
43.
by qrian 3y ago
Not a scientist but I learned those two concept from following the question of 'what the hell is a causality?' and reading related articles. My tentative conclusion is that we still haven't figure it out what causality is, an
44.
by qrian 3y ago
Author should also look into complexity science, and its consequences that even if everything is determined, most information about next state quickly evaporates. If it's impossible to know most things about next state, does it even ma
45.
by qrian 3y ago
Am I paranoid or is this a clear evidence that we cannot control nor contain AI as people will try to misuse it as much as possible for convenience and profit?
46.
by qrian 3y ago
It's interesting that they tested for AI code completion compatibility. It makes sense as AI code completion is now a big part of tooling and standards should accomodate for common tools.
47.
by qrian 3y ago
I vaguely remember this as a key concept in game design, however I cannot refer to it as it does not have a canonical name as far as I know. Does it have a canonical name?
48.
by qrian 3y ago
I'd also like to join. Care to open up the invitation link again?
49.
by qrian 3y ago
I've seen one determined by a harry potter quiz. > Order of authorship was determined by the top score each person received across all three levels of the official Pottermore “Back to Hogwarts Quiz.” For information about the exact
50.
by qrian 4y ago
Few weeks of intense reading to understand a whole semester's worth of mathematics... I'd also like to have that skill. Perhaps being homeschooled helped acquire such skill.
51.
by qrian 4y ago
Current discussions of AI is novel because they now show emergent behaviours. Reductionist view on humans are still incorrect, but what changed is that AI are viewed with non-reductionist views. Nobody takes Chomskian views of 'AIs wil
52.
by qrian 4y ago
Q) Player 1 has King at e8, and no other pieces. Player 2 has only King at e6 and Rook at h1. It is Player 2's move. What do Player 2 play? ChatGPT: Player 2 should move the Rook to h8 to checkmate Player 1.
53.
by qrian 4y ago
The chess notation used in the original article[1] is different from the notation we use today[2]. Might want to consider that. Though, the response that ChatGPT has given seems like the best possible answer if they didn't know about t
54.
by qrian 4y ago
Look for artifacts in hands, texts, and also look for incoherence among distant sections. Got 27 out of 30 in first try.
55.
by qrian 4y ago
I’m korean and I bet they are just initialisms for searching purposes. When I want to open “safari”, instead of typing the whole word (사파리), I can just type in the initials “ㅅㅍㄹ” and safari will be searched. I can see apple implementing s
56.
by qrian 4y ago
Looks like they just slapped 5$ pricing on the previous free tier, which also gave 1000 dyno hours with 30 minutes inactivity sleep.
57.
by qrian 4y ago
I work in a relevant field. AI powered personalized instructor was always around the corner like how autonomous car was around the corner since 2016. And we understand less about ‘understanding’ than driving cars.
58.
by qrian 4y ago
There's also [Quantum Smell Theory]( https://www.youtube.com/watch?v=DJsJIVXkrGQ ) where quantum tunneling is proposed as the primary mechanism for detecting the shape of a molecule being smelled.
59.
by qrian 4y ago
> I will note that the United States Postal Service has an employed senior official whose only job is making sure the largest banks in the U.S. don’t go paperless-by-default. This sounds interesting and plausable, but can anyone find a s
60.
by qrian 4y ago
"Hell Korea" is what koreans call their country, and also we're going kinda extinct.

More ›